On 04.02.25 13:53, Johannes Berg wrote:
On Tue, 2025-02-04 at 13:26 +0100, Alexander Wetzel wrote:Create/delete debugfs entry for the virtual monitor interface according to the state of the virtual monitor interface. This fixes the following issues: 1) Debugfs is not getting an entry when the virtual monitor interface is created. 2) Instead debugfs gets an entry when it's deleted. 3) And debugfs warns, that it already has directory for the virtual monitor interface every time the virtual monitor interface is removed again after the first time.Started wondering, maybe it should just not have debugfs at all? That'd also avoid the potential name clash, and there isn't really anything of value there in the first place?
Knowing that we do have an internal virtual monitor interface may be interesting in itself. But that's hardly something we can call "real value":-) And that way we would solve any name conflict...
Not creating it would just require us to skip ieee80211_debugfs_recreate_netdev() in drv_remove_interface() for virtual monitor.
When you prefer that, I can send an alternate patch doing that. Do you want that as standalone patch or as a v2 of with a new name? Alexander
